Behavior Design Sub-Skill

Version: 1.5.0
Framework: Jerry User-Experience -- Behavior Design
Constitutional Compliance: Jerry Constitution v1.0
Parent Skill: /user-experience (skills/user-experience/SKILL.md)
Wave: 4 (Advanced Analytics)
Project: PROJ-022 User Experience Skill | GitHub Issue #138

Document Sections

Section	Purpose
Document Audience	Triple-Lens audience guide
Purpose	Sub-skill overview and key capabilities
When to Use This Sub-Skill	Activation triggers and scope boundaries
Available Agents	Single agent with role, model, and output location
P-003 Compliance	Worker agent hierarchy position
Invoking the Agent	Invocation via ux-orchestrator
Methodology	Fogg B=MAP framework, factor analysis, bottleneck identification, intervention design, 5-phase execution procedure
Output Specification	Output location, L0/L1/L2 structure, required sections
Routing	Keywords and lifecycle-stage routing integration
Cross-Framework Integration	Handoff from heuristic evaluation and to HEART metrics
Synthesis Hypothesis Confidence	Confidence classifications for Behavior Design outputs
Quality Gate Integration	S-014 scoring and H-13 threshold enforcement
Degraded Mode Behavior	Operation without real-time behavioral data
Wave Architecture	Wave 4 entry criteria, bypass conditions
Constitutional Compliance	Governing principles and AI-augmented analysis limitations
Registration	H-26 parent-routed registration model and AGENTS.md confirmation
Deployment Status	Wave 4 stub agent status and implementation timeline
Quick Reference	Common workflows and agent selection hints
References	Full repo-relative paths, requirements traceability, external citations

Document Audience (Triple-Lens)

This SKILL.md serves multiple audiences:

Level	Audience	Sections to Focus On
L0 (Stakeholder)	Product managers, designers	Purpose, When to Use This Sub-Skill, Quick Reference
L1 (Developer)	Engineers invoking the agent	Invoking the Agent, Methodology, Output Specification
L2 (Architect)	Workflow designers, skill maintainers	Cross-Framework Integration, Synthesis Hypothesis Confidence, Degraded Mode Behavior

---

Purpose

The Behavior Design sub-skill provides structured behavioral bottleneck diagnosis using BJ Fogg's Behavior Model, commonly expressed as B=MAP: Behavior happens when Motivation, Ability, and a Prompt converge at the same moment (Fogg, 2009; Fogg, 2020). It targets tiny teams (1-5 people) who observe users failing to complete desired actions and need a systematic framework to identify which behavioral factor is the limiting constraint.

This sub-skill is part of Wave 4 (Advanced Analytics), requiring Wave 3 completion before deployment. It bridges design system construction (Wave 3) and process-intensive activities (Wave 5) by providing behavioral insight that explains why well-designed interfaces still fail to drive target user actions.

Key Capabilities

• B=MAP Factor Assessment -- Systematically evaluates each of the three behavioral factors (Motivation, Ability, Prompt) for a target behavior, determining whether each factor is above or below the action threshold (Fogg, 2020)
• Motivation Analysis -- Analyzes intrinsic motivators (sensation, anticipation, belonging), extrinsic motivators (rewards, punishments, social proof), and social drivers (competition, collaboration, recognition) that influence user willingness to act
• Ability Analysis -- Evaluates the six Fogg simplicity factors (Time, Money, Physical Effort, Brain Cycles, Social Deviance, Non-Routine) to identify friction points that make the target behavior too difficult (Fogg, 2009)
• Prompt Analysis -- Classifies prompts into three types (Facilitator, Signal, Spark) and assesses whether the right prompt type is deployed at the right moment for the user's motivation-ability state (Fogg, 2009)
• Bottleneck Identification -- Determines which single factor (or combination) is the primary constraint preventing the target behavior, using a structured elimination algorithm
• Intervention Design -- Recommends targeted interventions that address the diagnosed bottleneck: motivation interventions for low-motivation bottlenecks, simplification for low-ability bottlenecks, or prompt redesign for missing/mistimed prompts

---

When to Use This Sub-Skill

Activate when:

• Users are not completing a specific desired action and the team needs to understand why
• Diagnosing whether user inaction stems from insufficient motivation, excessive difficulty, or absent/mistimed prompts
• Analyzing post-launch behavioral data showing low conversion, abandonment, or incomplete task flows
• Designing behavior change interventions targeted at a specific bottleneck factor
• Investigating why a well-designed interface fails to drive the intended user action
• Providing behavioral root-cause analysis for urgent UX problems (CRISIS mode step 2)
• Evaluating whether a proposed feature change addresses the correct behavioral bottleneck
• Preparing behavioral diagnosis to feed into HEART metrics measurement

Do NOT use for:

• Evaluating an existing interface against usability heuristics -- use /ux-heuristic-eval (Nielsen's 10) instead. Use heuristic evaluation first, then Behavior Design to trace severe issues to behavioral root causes.
• Building or auditing a component library -- use /ux-atomic-design (Atomic Design) instead.
• Accessibility compliance auditing -- use /ux-inclusive-design (WCAG 2.2) instead.
• Measuring quantitative UX health metrics -- use /ux-heart-metrics (Google GSM) instead. Use Behavior Design first to diagnose, then HEART to measure improvement.
• Understanding user motivations at the job level -- use /ux-jtbd (Jobs-to-Be-Done) instead. JTBD identifies what users want; Behavior Design diagnoses why they fail to complete a specific action.
• Testing hypotheses about design changes -- use /ux-lean-ux (Lean UX) instead.
• Running a full rapid prototyping sprint -- use /ux-design-sprint (Design Sprint 2.0) instead.
• Prioritizing features by user satisfaction impact -- use /ux-kano-model (Kano) instead.
• Security-focused interface review -- use /eng-team instead.
• General research without behavioral UX focus -- use /problem-solving instead.

---

Available Agents

Agent	Role	Tier	Mode	Model	Output Location
ux-behavior-diagnostician	Fogg B=MAP behavior bottleneck diagnostician	T2	Convergent	Sonnet	skills/ux-behavior-design/output/{engagement-id}/ux-behavior-diagnostician-{topic-slug}.md

STUB: The agent definition file (skills/ux-behavior-design/agents/ux-behavior-diagnostician.md) is pending Wave 4 Phase 2 implementation as part of PROJ-022 EPIC-004. The SKILL.md specifies the methodology and output contract that the agent will implement.

Tool tier: T2 (Read-Write) = Read, Write, Edit, Glob, Grep, Bash. No WebSearch, WebFetch, or Context7 MCP access -- the B=MAP methodology is self-contained. See .context/rules/agent-development-standards.md [Tool Security Tiers].

The agent produces output at three levels per AD-M-004:

• L0 (Executive Summary): Primary bottleneck identification (Motivation, Ability, or Prompt); severity assessment; top intervention recommendation; key findings for stakeholders and cross-framework synthesis input.
• L1 (Technical Detail): Full B=MAP factor assessment with evidence, simplicity factor breakdown, prompt type classification, bottleneck elimination algorithm trace, and detailed intervention recommendations.
• L2 (Strategic Implications): Behavioral pattern analysis across multiple target behaviors, systemic bottleneck trends, organizational behavior design maturity assessment, and behavior change roadmap.

---

P-003 Compliance

The /ux-behavior-design sub-skill contains a single worker agent. It is invoked by the ux-orchestrator (T5) via the Task tool. The agent does NOT have Task tool access and MUST NOT spawn sub-agents.

MAIN CONTEXT (user request)
    |
    v
ux-orchestrator (T5, Opus, Integrative) -- parent orchestrator
    |
    +-- ux-behavior-diagnostician (T2, Convergent, Sonnet) -- THIS sub-skill's worker
    +-- [other sub-skill workers...]

Enforcement:

• disallowedTools: [Task] declared in skills/ux-behavior-design/agents/ux-behavior-diagnostician.md frontmatter
• P-003 prohibition in skills/ux-behavior-design/agents/ux-behavior-diagnostician.governance.yaml capabilities.forbidden_actions
• CI gate validates no sub-skill agent has Task access (documented in skills/user-experience/rules/ci-checks.md)

---

Invoking the Agent

This is a sub-skill invoked by the ux-orchestrator, not directly by users. Users interact with the parent /user-experience skill, which routes to this sub-skill based on lifecycle-stage triage.

Via Natural Language (to parent skill)

"Why aren't users completing the checkout process?"
"Diagnose the behavioral bottleneck in our onboarding flow"
"Users see the CTA but don't click it -- what's blocking them?"
"Analyze motivation, ability, and prompts for the signup action"
"Why are users abandoning the form after filling out 3 of 5 fields?"
"What's preventing users from upgrading to the paid plan?"

Via Explicit Agent Request (to parent skill)

"Use ux-behavior-diagnostician to diagnose why users abandon the payment form"
"Have ux-behavior-diagnostician analyze B=MAP factors for the onboarding tutorial"

Via Task Tool (orchestrator internal)

The ux-orchestrator invokes the agent via the Task tool:

Task(
    description="ux-behavior-diagnostician: B=MAP bottleneck diagnosis for checkout abandonment",
    subagent_type="jerry:ux-behavior-diagnostician",
    prompt="""
## UX CONTEXT (REQUIRED)
- **Engagement ID:** UX-0001
- **Topic:** Checkout Abandonment Bottleneck Diagnosis
- **Product:** [product name and domain]
- **Target Users:** [user description]
- **Target Behavior:** [specific action users should take but are not taking]
- **Current Behavior Data:** [what users are doing instead, abandonment rates, funnel data]
- **Input:** [screenshots, flow descriptions, upstream heuristic findings]

## TASK
Diagnose the behavioral bottleneck preventing users from completing checkout.
1. Define the target behavior in specific, observable terms
2. Assess Motivation factors (intrinsic, extrinsic, social drivers)
3. Assess Ability factors (6 Fogg simplicity factors)
4. Assess Prompt factors (type, timing, placement)
5. Identify the primary bottleneck via elimination algorithm
6. Recommend targeted interventions for the diagnosed bottleneck

## MANDATORY PERSISTENCE (P-002)
Create file at: skills/ux-behavior-design/output/UX-0001/ux-behavior-diagnostician-checkout-abandonment.md
"""
)

Governance codification (AD-M-007): The session_context contract (on_receive/on_send) is specified in ux-behavior-diagnostician.governance.yaml per AD-M-007. Fields are enumerated below:

on_receive fields:

Field	Type	Required	Description
engagement_id	string	Yes	UX engagement identifier (format: UX-{NNNN})
product_context	string	Yes	Product name, domain, and target user description
target_behavior	string	Yes	Specific, observable action users should take but are not taking (e.g., "click 'Complete Purchase' button on checkout page")
current_behavior_data	string	No	Observed user behavior: abandonment rates, funnel stage, click data, session recordings summary
upstream_artifacts	array	No	File paths to upstream handoff artifacts (heuristic evaluation severity-rated findings, screenshots, flow descriptions)

on_send fields:

Field	Type	Required	Description
bottleneck_factor	enum	Yes	Primary bottleneck: motivation, ability, prompt, or multiple
bottleneck_severity	enum	Yes	critical (never occurs), major (rare), moderate (inconsistent)
motivation_assessment	object	Yes	Intrinsic/extrinsic/social scores (1-5 each), overall above/below threshold
ability_assessment	object	Yes	Six simplicity factor scores (1-5 each), overall above/below threshold
prompt_assessment	object	Yes	Type (facilitator/signal/spark), timing, placement
interventions	array	Yes	Interventions with effort estimate and expected impact
synthesis_judgments	array	Yes	AI judgment calls with confidence classification and rationale

---

Methodology

Fogg Behavior Model: B=MAP

The Fogg Behavior Model (Fogg, 2009, DOI: 10.1145/1541948.1541999; Fogg, 2020) states that B=MAP: behavior occurs when Motivation, Ability, and Prompt converge above their respective thresholds at the same moment. If any factor is missing or below the action threshold, the behavior does not occur -- all three factors must be simultaneously sufficient.

The model uses a motivation (Y-axis) vs. ability (X-axis) plane with a curved action line. A user with high motivation but high friction falls below the line; a user who finds the action easy but lacks motivation also falls below. The prompt must arrive when the user is above the action line (Fogg, 2009).

Motivation Analysis

Motivation in the Fogg model operates through three core motivator pairs (Fogg, 2009):

Motivator Pair	High End	Low End	Assessment Question
Sensation	Pleasure	Pain	Does the action produce immediate pleasure or relieve pain?
Anticipation	Hope	Fear	Does the user hope for reward or fear consequence?
Belonging	Acceptance	Rejection	Does the action increase social standing or prevent exclusion?

Intrinsic motivators: Internal drives (satisfaction, curiosity, mastery, autonomy). Durable but difficult to create through design alone (Fogg, 2020).

Extrinsic motivators: External incentives (rewards, punishments, progress indicators). Easier to design but less durable; decay when the incentive is removed (Fogg, 2020).

Social drivers: Peer influence (competition, collaboration, recognition, social proof). Operate through the belonging motivator pair (Fogg, 2009).

Motivation assessment scale: Each motivation dimension (intrinsic, extrinsic, social) is rated 1-5 based on available evidence:

• 1-2: Below threshold (motivation is likely the bottleneck)
• 3: At threshold (borderline; investigate further)
• 4-5: Above threshold (motivation is not the primary bottleneck)

Ability Analysis: Six Simplicity Factors

Ability in the Fogg model is inversely related to simplicity. The six simplicity factors (Fogg, 2009) represent distinct dimensions of friction that can prevent a user from completing the target behavior:

Factor	Definition	Assessment Question	Example Friction
Time	How long the behavior takes to complete	Does the action take more time than the user is willing to spend?	12-field form when user expects 3 fields
Money	Financial cost of the behavior	Does the action require spending more money than the user expected?	Surprise shipping costs at checkout
Physical Effort	Bodily exertion required	Does the action require more physical effort than the user is willing to expend?	Requiring camera access, photo upload, manual data entry on mobile
Brain Cycles	Cognitive load required	Does the action require more thinking than the user is willing to invest?	Complex pricing tiers requiring comparison; unfamiliar terminology
Social Deviance	Degree to which the behavior violates social norms	Does the action require the user to do something socially unacceptable in their context?	Publicly sharing salary information; requesting social media permissions
Non-Routine	Degree to which the behavior departs from the user's established habits	Does the action require the user to change an existing habit or adopt a new workflow?	Switching from email-based workflow to an unfamiliar dashboard

Simplicity factor assessment scale: Each factor is rated 1-5:

• 1-2: High friction (this factor is likely contributing to the bottleneck)
• 3: Moderate friction (borderline; warrants investigation)
• 4-5: Low friction (this factor is not the primary obstacle)

The limiting simplicity factor is the factor with the lowest score. Per Fogg (2009), ability is governed by the scarcest resource at the moment of the prompt -- improving any factor except the limiting one does not raise overall ability above the action threshold.

Prompt Analysis

Prompts are the triggers that call users to action. The Fogg model identifies three prompt types, each appropriate for a specific motivation-ability state (Fogg, 2009):

Prompt Type	User State	Mechanism	Design Implication
Spark	High ability, low motivation	The prompt simultaneously triggers motivation (e.g., emotional appeal, social proof, fear of missing out)	User can do the action but needs a reason; the prompt provides that reason
Facilitator	High motivation, low ability	The prompt simultaneously reduces friction (e.g., pre-filled forms, one-click actions, simplified steps)	User wants to act but the action is too hard; the prompt makes it easier
Signal	High motivation, high ability	The prompt serves as a simple reminder or notification (e.g., bell icon, push notification, calendar reminder)	User is both willing and able; they just need a timely nudge

Prompt assessment dimensions:

Dimension	Assessment	Values
Type match	Is the prompt type appropriate for the user's motivation-ability state?	Appropriate (correct type for state), Mismatched (wrong type -- e.g., signal prompt when facilitator is needed)
Timing	Does the prompt arrive when the user is ready to act?	Appropriate (arrives at the right moment), Mistimed (arrives too early or too late), Absent (no prompt exists)
Placement	Is the prompt visible and actionable when it fires?	Visible (user notices it), Hidden (below fold, in a submenu), Competing (surrounded by other CTAs that dilute attention)

Bottleneck Identification Algorithm

The diagnostician follows a structured elimination algorithm to identify the primary bottleneck:

Step 1: Prompt Assessment (cheapest fix first)

• Prompt absent, mistimed, or mismatched -> prompt is the primary bottleneck. A perfect motivation-ability combination produces zero behavior without a prompt (Fogg, 2020).
• Prompt present, timed, and type-appropriate -> proceed to Step 2.

Step 2: Ability Assessment (most common bottleneck)

• Calculate the limiting simplicity factor (lowest-scoring among six).
• Limiting factor scores 1-2 -> ability is the primary bottleneck. Ability is the most common bottleneck in digital products (Fogg, 2020).
• All factors score 3+ -> proceed to Step 3.

Step 3: Motivation Assessment (hardest to change)

• Majority of motivation dimensions score 1-2 -> motivation is the primary bottleneck.
• Motivation above threshold -> proceed to Step 4.

Step 4: Multiple Bottleneck Assessment

• Two or more factors at borderline (score 3) with none clearly below threshold -> classify as multiple bottleneck requiring coordinated interventions.

Algorithm ordering rationale: The prompt-first, ability-second, motivation-third ordering follows Fogg's (2020) intervention difficulty gradient: prompts are cheapest to fix, ability is moderate, motivation is hardest.

Intervention Design Recommendations

Based on the diagnosed bottleneck, the diagnostician recommends interventions targeted at the weakest factor:

Bottleneck	Intervention Category	Example Interventions	Effort
Prompt	Add, reposition, or retype the prompt	Add CTA where none exists; move above fold; change signal to spark for low-motivation users	Low
Ability (Time)	Reduce time required	Pre-fill fields; reduce steps; add progress indicators; offer express path	Low-Medium
Ability (Money)	Reduce or clarify cost	Show total upfront; offer free tier; defer payment	Medium
Ability (Physical Effort)	Reduce physical actions	Add autofill; one-click actions; optimize for mobile	Low-Medium
Ability (Brain Cycles)	Reduce cognitive load	Simplify language; add defaults; reduce choices; add tooltips	Medium
Ability (Social Deviance)	Normalize the behavior	Add social proof; anonymization options; privacy controls	Medium
Ability (Non-Routine)	Reduce habit disruption	Map to existing workflows; familiar UI patterns; transition guides	Medium-High
Motivation	Increase motivation via design	Social proof; gamification; loss aversion framing; progress visualization	High
Multiple	Coordinated multi-factor	Address lowest-scoring factor first; combine prompt redesign with simplification	High

Execution Phases

Note: This execution procedure describes target behavior for the fully-implemented ux-behavior-diagnostician agent. The current agent definition is a Wave 4 stub; full implementation will follow this specification.

The diagnostician follows a 5-phase sequential workflow. Each phase produces intermediate artifacts that feed the next. This mirrors the Phase 1-5 structure established by the HEART metrics, Lean UX, and Atomic Design sub-skills.

Phase 1: Scope Definition

Purpose: Establish the behavioral context, define the target behavior, confirm wave entry criteria, and catalog available behavioral evidence.

Activities:

1. Identify the product domain, target users, and the specific action users should take but are not taking. Define the target behavior using Fogg's statement format: "After [CONTEXT], I will [SPECIFIC BEHAVIOR]" (Fogg, 2020, Chapters 14-15)
2. Confirm Wave 4 entry criteria are met: Wave 3 completed (Storybook with 5+ Atom stories AND 1 Persona Spectrum review), OR bypass condition satisfied (existing user base with analytics). (Verification: check for WAVE-3-SIGNOFF.md in skills/user-experience/output/; if absent, ask user per H-31.)
3. Catalog upstream inputs: check for /ux-heuristic-eval severity-rated findings (severity >= 2); if present, import finding IDs for bottleneck context
4. Catalog available behavioral evidence: abandonment rates, funnel data, session recordings, support tickets, interview excerpts. Classify as strong (quantitative), moderate (qualitative), or weak (anecdotal)
5. Establish observation scope: screens, flows, or interaction sequences containing the target behavior; current conversion rate (if known)

Output: Scope brief with the following required fields:

Field	Description	Example
Product Domain	Application area and target user segment	"E-commerce checkout flow for first-time buyers"
Target Behavior Statement	Fogg format: "After [CONTEXT], I will [SPECIFIC BEHAVIOR]"	"After adding items to cart, I will complete purchase"
Observation Scope	Screens, flows, or interaction sequences under analysis	"Cart page through order confirmation (4 screens)"
Upstream Findings	Imported heuristic evaluation findings (if available), with finding IDs	"HE-003 (severity 3): unclear CTA placement"
Evidence Inventory	Available behavioral evidence classified as strong/moderate/weak	"Strong: 23% cart abandonment rate; Moderate: 3 user interviews; Weak: support ticket anecdotes"
Wave Entry Status	Wave 4 criteria verification result	"PASS: WAVE-3-SIGNOFF.md present, 7 Atom stories confirmed"

Phase 2: Behavior Mapping

Purpose: Map the current state of each B=MAP factor, establishing a baseline assessment.

Activities:

1. Assess Motivation: evaluate each dimension (intrinsic, extrinsic, social) on the 1-5 scale. Cite specific evidence; mark inferences where no direct evidence exists.
2. Assess Ability: evaluate all six simplicity factors (1-5 each). Describe friction points or their absence. Identify the limiting simplicity factor.
3. Assess Prompts: classify by type (Facilitator/Signal/Spark), evaluate timing (appropriate/mistimed/absent) and placement (visible/hidden/competing). Record "prompt absent" if none exists.
4. Plot the user's approximate position on the motivation-ability plane. Describe whether the user is in the action zone and which factor(s) need to change.

Output: B=MAP state map: motivation scores, ability scores with limiting factor, prompt assessment, action-line position.

Phase 3: Bottleneck Diagnosis

Purpose: Apply the bottleneck identification algorithm to determine the primary constraint.

Activities:

1. Execute Step 1 (Prompt Assessment): evaluate prompt presence, type, timing, placement. If any criterion fails, classify as prompt bottleneck.
2. Execute Step 2 (Ability Assessment): if prompt is adequate, examine the limiting simplicity factor. If it scores 1-2, classify as ability bottleneck with the specific factor identified and evidence chain.
3. Execute Step 3 (Motivation Assessment): if ability is adequate, evaluate overall motivation. If majority of dimensions score 1-2, classify as motivation bottleneck with evidence.
4. Execute Step 4 (Multiple Bottleneck Assessment): if two or more factors are borderline (score 3), classify as multiple bottleneck.
5. Assign bottleneck severity: critical (zero conversion), major (conversion below 10% of expected), moderate (conversion 10-50% of expected). (Heuristic thresholds: 10% and 50% are framework-internal heuristics. Adjust based on domain-specific baselines.)

Output: Bottleneck diagnosis: primary factor, severity, evidence chain, algorithm trace, confidence assessment.

Phase 4: Intervention Design

Purpose: Recommend targeted interventions addressing the diagnosed bottleneck.

Activities:

1. Select intervention category based on diagnosed bottleneck (see Intervention Design Recommendations table). For ability bottlenecks, target the specific limiting simplicity factor.
2. Generate 3-5 specific interventions, each with: description, target factor, expected impact (high/medium/low), implementation effort (low/medium/high), and supporting reasoning.
3. Prioritize by effort-to-impact ratio: low-effort, high-impact first. Prompt and ability interventions before motivation interventions (Fogg, 2020).
4. Classify each intervention as direct (addresses primary bottleneck) or supporting (reinforces primary intervention).
5. Mark all intervention recommendations with LOW synthesis confidence per Synthesis Hypothesis Confidence.

Output: Prioritized interventions: description, target factor, impact, effort, direct/supporting, confidence (all LOW).

Phase 5: Synthesis and Handoff Preparation

Purpose: Synthesize findings, produce L0/L1/L2 output artifact, construct downstream handoff.

Activities:

1. Produce L0 executive summary: primary bottleneck, severity, top intervention, key findings
2. Produce L1 technical detail: full B=MAP assessment, algorithm trace, intervention list, evidence citations
3. Produce L2 strategic implications: behavioral pattern analysis, systemic trends, maturity assessment, behavior change roadmap
4. Compile Synthesis Judgments Summary: every AI judgment call with confidence classification and rationale
5. Prepare /ux-heart-metrics handoff: bottleneck diagnosis with HEART dimension mapping for measurement baseline
6. If CRISIS mode: add priority ranking and quick-win identification (prompt bottlenecks flagged as quick wins per Fogg, 2020)

Output: Complete output artifact per the Required Output Sections specification (L0 executive summary, L1 technical sections, L2 strategic implications, synthesis judgments, handoff data). Handoff payload for /ux-heart-metrics.

---

Output Specification

Output Location

skills/ux-behavior-design/output/{engagement-id}/ux-behavior-diagnostician-{topic-slug}.md

Where:

• {engagement-id} follows the pattern UX-{NNNN} (e.g., UX-0001)
• {topic-slug} is a kebab-case descriptor of the target behavior (e.g., checkout-abandonment, onboarding-completion, plan-upgrade)

Required Output Sections

Section	Level	Content
Executive Summary	L0	Primary bottleneck factor; bottleneck severity; top intervention recommendation; key findings for stakeholders and cross-framework synthesis input
Engagement Context	L1	Product description, target users, target behavior statement, available evidence inventory, observation scope, upstream inputs, wave entry verification
Behavior State Map	L1	Full B=MAP assessment: motivation scores (intrinsic, extrinsic, social; 1-5 each), ability scores (6 simplicity factors; 1-5 each with limiting factor identified), prompt assessment (type, timing, placement), action-line position
Bottleneck Diagnosis	L1	Primary bottleneck factor with elimination algorithm trace; bottleneck severity; evidence chain; confidence assessment
Intervention Recommendations	L1	Prioritized interventions (3-5) with: description, target factor, expected impact, implementation effort, direct/supporting classification; all marked LOW confidence
Strategic Implications	L2	Behavioral pattern analysis; systemic bottleneck trends; behavior design maturity assessment; behavior change roadmap
Synthesis Judgments Summary	L1	Each AI judgment call listed for synthesis confidence gate compliance
Handoff Data	L1	Structured data for downstream sub-skills: bottleneck diagnosis with HEART metric mapping (for /ux-heart-metrics measurement baseline)

Synthesis Judgments Summary requirements: MUST list every AI judgment (motivation ratings, simplicity scores, bottleneck classification, interventions) with confidence classification (HIGH/MEDIUM/LOW) and rationale. Each judgment row includes: finding ID, framework source (e.g., B=MAP factor, simplicity factor, intervention category), confidence level (HIGH/MEDIUM/LOW), and rationale explaining the classification basis and evidence chain. Follows the pattern in skills/user-experience/rules/synthesis-validation.md [STUB: EPIC-001].

Templates

Template	Path	Purpose
B=MAP Diagnosis Template	skills/ux-behavior-design/templates/bmap-diagnosis-template.md	B=MAP factor assessment with scoring tables, bottleneck algorithm trace, and intervention recommendation format

---

Routing

Trigger Keywords

Keyword	Routing Context
behavior design	Direct match -- primary trigger
B=MAP	Direct match
Fogg model	Direct match
behavior bottleneck	Direct match
motivation analysis	In combination with UX/behavior context
ability analysis	In combination with UX/behavior context
prompt design	In combination with behavior/UX context (not general prompt engineering)
why users don't	Direct match (phrase)
user inaction	Direct match
behavior diagnosis	Direct match
tiny habits	In combination with UX/design context
action threshold	In combination with behavior/UX context

Lifecycle-Stage Routing Integration

This sub-skill is routed to by the ux-orchestrator in the following lifecycle-stage scenarios:

Stage	User Intent	Route Condition
After launch	"Users not completing action"	Direct route to /ux-behavior-design; source: skills/user-experience/rules/ux-routing-rules.md Stage Routing Table (pending EPIC-001 completion)
After launch	Follow-up from heuristic evaluation	When /ux-heuristic-eval has identified severity >= 2 findings with behavioral implications; heuristic findings provide context for bottleneck diagnosis
CRISIS	Urgent UX problems (step 2)	CRISIS mode invokes Behavior Design as step 2 of the fixed 3-skill emergency sequence (Heuristic Eval -> Behavior Design -> HEART Metrics); source: skills/user-experience/rules/ux-routing-rules.md CRISIS Routing (pending EPIC-001 completion)
Any stage	"Fix a specific UX problem" (behavioral)	Qualification question: "Is the problem about user behavior or design quality?" -> Behavior: /ux-behavior-design; source: skills/user-experience/rules/ux-routing-rules.md Common Intent Resolution (pending EPIC-001 completion)

Wave Gating

This sub-skill is in Wave 4 (Advanced Analytics). It requires Wave 3 completion before deployment:

Entry criteria: Wave 3: Storybook with 5+ Atom stories AND 1 Persona Spectrum review.

Bypass condition: Existing user base with analytics (skip Persona Spectrum prerequisite). This bypass recognizes that teams with an existing user base and analytics data have the behavioral evidence needed for B=MAP diagnosis, even without the Persona Spectrum review that Wave 3 inclusive design would normally provide.

---

Cross-Framework Integration

Upstream Inputs

This sub-skill receives context from other sub-skills when invoked as part of a multi-sub-skill workflow:

From Sub-Skill	Handoff Artifact	Key Fields	Usage
/ux-heuristic-eval	Severity-rated findings	Finding ID, heuristic violated, severity (0-4), affected screen/flow	Heuristic findings with severity >= 2 provide context for behavioral diagnosis; high-severity findings identify specific UI locations where user behavior breaks down, guiding the B=MAP scope definition

Downstream Handoffs

This sub-skill produces artifacts that feed into other sub-skills via the Jerry handoff protocol (docs/schemas/handoff-v2.schema.json).

To Sub-Skill	Handoff Artifact	Key Fields	Trigger
/ux-heart-metrics	Bottleneck diagnosis with HEART metric mapping	Bottleneck factor, bottleneck severity, affected screen/flow, candidate HEART metric category (Happiness/Engagement/Adoption/Retention/Task Success), intervention list	After bottleneck diagnosis is complete; HEART analyst establishes measurement baselines for the diagnosed bottleneck area

Handoff data format (handoff-v2 + ux-ext):

handoff:
  from_agent: ux-behavior-diagnostician
  to_agent: ux-heart-analyst
  task: "Establish HEART metric baselines for diagnosed behavioral bottleneck"
  success_criteria:
    - "Metric baselines established for affected HEART dimension"
    - "Target thresholds set for post-intervention measurement"
  artifacts:
    - "skills/ux-behavior-design/output/{engagement-id}/ux-behavior-diagnostician-{topic-slug}.md"
  key_findings:
    - "Primary bottleneck: {factor} ({severity})"
    - "Limiting simplicity factor: {factor_name} (score: {N}/5)"
  blockers: []
  confidence: 0.6
  criticality: C2
  ux_ext:
    bottleneck_factor: "{motivation|ability|prompt|multiple}"
    bottleneck_severity: "{critical|major|moderate}"
    affected_heart_dimension: "{happiness|engagement|adoption|retention|task_success}"

Canonical Multi-Skill Workflow Sequences

This sub-skill participates in the following canonical sequences:

Sequence	Skills Involved	This Sub-Skill's Role
Evaluate to Diagnose to Measure	/ux-heuristic-eval then /ux-behavior-design then /ux-heart-metrics	Receives heuristic findings and traces them to behavioral root causes via B=MAP; hands off bottleneck diagnosis to HEART for measurement
CRISIS Emergency Sequence	/ux-heuristic-eval then /ux-behavior-design then /ux-heart-metrics	Step 2 of the fixed CRISIS 3-skill sequence; diagnoses behavioral root causes of urgent UX problems

---

Synthesis Hypothesis Confidence

Behavior Design outputs include synthesis hypotheses that carry confidence classifications per the synthesis validation protocol.

Synthesis Step	Typical Confidence	Rationale
B=MAP bottleneck diagnosis	MEDIUM	Fogg Behavior Model (Fogg, 2020) provides structured diagnosis but bottleneck attribution requires user-specific data. Framework constrains to three factors, reducing interpretive variance, but scores depend on evidence quality.
Design intervention recommendation	LOW	Intervention effectiveness depends on context-specific factors (demographics, domain, constraints) that training data cannot capture. Directionally sound per B=MAP but requires user testing (Fogg, 2020).

Gate enforcement:

• MEDIUM outputs (bottleneck diagnosis): Include "Validation Required" section. Recommendations conditional on validation against 2-3 real user data points (interviews, analytics, or session recordings).
• LOW outputs (intervention recommendations): Intervention section tagged [REFERENCE-ONLY]. Banner: "Intervention recommendations are directional based on B=MAP analysis. Effectiveness requires validation through user testing or A/B experimentation."

Note on confidence dynamics: Bottleneck diagnosis can achieve HIGH synthesis confidence when converged with a second framework -- for example, when /ux-heuristic-eval identifies severity >= 3 findings corroborating the diagnosed bottleneck. Intervention recommendations remain LOW regardless of diagnostic confidence because effectiveness requires empirical testing.

---

Quality Gate Integration

Behavior Design outputs are subject to the Jerry quality gate per H-13 and H-14:

Quality Check	Threshold	Application
S-014 LLM-as-Judge scoring	>= 0.92 composite (C2+)	Applied at Phase 5 completion
Creator-critic-revision	Minimum 3 iterations (H-14)	Orchestrator manages revision cycles
Self-review (S-010)	Required before presenting	Self-review before returning to orchestrator
Wave transition gate	S-014 composite >= 0.85	Applied at Wave 4 -> 5 transition

Scoring dimensions (Behavior Design interpretation):

Dimension	Weight	Interpretation
Completeness	0.20	All B=MAP factors assessed; all six simplicity factors rated; prompt analysis included
Internal Consistency	0.20	Bottleneck diagnosis consistent with factor scores; interventions target diagnosed bottleneck
Methodological Rigor	0.20	Fogg model correctly applied; elimination algorithm followed; factor ratings evidence-based
Evidence Quality	0.15	Factor ratings cite evidence; evidence quality classified; inferences labeled
Actionability	0.15	Interventions specific, implementable, prioritized by effort-to-impact
Traceability	0.10	Findings trace to evidence; synthesis judgments documented; upstream findings cited

CI Gate Summary

The following CI gate criteria apply to this sub-skill (full gate definitions in skills/user-experience/rules/ci-checks.md):

Gate	Check	Enforcement
No Task tool access	disallowedTools: [Task] present in agent frontmatter; agent MUST NOT have Task in tools list	L5 (CI): grep agent frontmatter for Task tool presence
P-003 forbidden action	capabilities.forbidden_actions in .governance.yaml MUST include P-003 recursive subagent prohibition	L5 (CI): schema validation of governance YAML against docs/schemas/agent-governance-v1.schema.json
Output schema validation	Agent output MUST contain all Required Output Sections (Executive Summary, Engagement Context, Behavior State Map, Bottleneck Diagnosis, Intervention Recommendations, Strategic Implications, Synthesis Judgments Summary, Handoff Data)	L4 (post-tool): section heading presence check on output artifact

---

Degraded Mode Behavior

The ux-behavior-diagnostician operates at T2 (Read-Write) and does NOT have access to WebSearch, WebFetch, or Context7 MCP tools. Additionally, it operates on user-provided descriptions and artifacts rather than live behavioral analytics. The following degraded modes apply:

No Real-Time Behavioral Data

When the user cannot provide quantitative behavioral data (analytics, funnel metrics, session recordings):

Limitation	Impact	Mitigation
No conversion rate data	Cannot quantify bottleneck severity precisely	Ask for qualitative assessment: "How often does this fail?" Map: "never" = critical, "rarely" = major, "sometimes" = moderate
No funnel analytics	Cannot identify exact drop-off point	Ask user to describe the last step before abandonment
No session recordings	Cannot observe actual behavior patterns	Ask: (1) "What do users do instead of the target action?" (2) "At what step do users stop? Describe the last thing they do before abandoning." (3) "Have you observed any user confusion or frustration signals (support tickets, rage clicks, dead-end navigation)?"
No A/B test data	Cannot validate interventions empirically	Mark all interventions LOW confidence; recommend user testing

Degraded mode disclosure (P-022):

[DEGRADED MODE] This output was produced without real-time behavioral analytics data.
Factor assessments are based on user-provided descriptions and available interface artifacts.
Limitations:
- Bottleneck severity estimated from qualitative descriptions, not quantitative data
- Ability factor scores may not reflect actual user friction without session recordings
- Intervention recommendations are directional and require empirical validation

No Upstream Heuristic Evaluation

When invoked without prior /ux-heuristic-eval output (e.g., direct invocation rather than CRISIS sequence), the diagnostician performs its own high-level interface assessment based on provided screenshots or descriptions. This is less rigorous than a formal heuristic evaluation. Bottleneck diagnosis references screen locations and behaviors rather than heuristic finding IDs.

No Screenshots or Design Artifacts

When the user provides only text descriptions without visual references, the diagnostician asks structured questions to compensate: prompt visibility and placement ("Where does the CTA appear? What else is visible around it?") and flow structure ("Describe the step sequence from entry to target action").

---

Wave Architecture

Wave 4: Advanced Analytics

This sub-skill is part of Wave 4 (Advanced Analytics), alongside /ux-kano-model.

Entry criteria: Wave 3 completed -- Storybook with 5+ Atom stories AND 1 Persona Spectrum review.

Bypass condition: Existing user base with analytics (skip Persona Spectrum prerequisite). Bypass requires 3-field documentation: unmet criterion, impact assessment ("behavior diagnosis proceeds without inclusive design context"), and remediation plan with target date.

Wave transition to Wave 5: Requires 30+ users for Kano survey OR 1 B=MAP bottleneck diagnosed.

---

Constitutional Compliance

All agents in this sub-skill adhere to the Jerry Constitution v1.0:

Principle	Requirement	Consequence of Violation
P-003	NEVER spawn recursive subagents -- worker agent, no Task tool access	Agent hierarchy violation; uncontrolled token consumption
P-020	NEVER override user decisions on bottleneck classification or intervention priorities	Unauthorized action; trust erosion
P-022	NEVER present bottleneck diagnoses as certain without disclosing evidence limitations; NEVER inflate factor scores without supporting evidence; NEVER omit the degraded mode disclosure when operating without behavioral data	Governance undermined; quality assessment invalidated
P-001	NEVER present factor ratings or bottleneck classifications without citing the evidence or reasoning supporting the assessment	Unreliable outputs; unfounded claims propagate downstream
P-002	NEVER leave bottleneck diagnoses or intervention recommendations in transient context only -- persist to files	Context rot vulnerability; artifacts lost on session compaction

Per-agent enforcement: The ux-behavior-diagnostician agent declares:

• constitution.principles_applied: P-003, P-020, P-022, P-001, P-002 in skills/ux-behavior-design/agents/ux-behavior-diagnostician.governance.yaml
• capabilities.forbidden_actions: 3 entries in NPT-009 format referencing the constitutional triplet
• disallowedTools: [Task] in skills/ux-behavior-design/agents/ux-behavior-diagnostician.md frontmatter

AI-Augmented Analysis Limitations

The Behavior Design diagnostician agent operates as an AI-augmented analysis tool. The following limitations apply to all outputs and MUST be disclosed per P-022 (no deception):

• Bottleneck diagnosis is interpretive. Factor assessment requires judgment. Different analysts may rate the same factors differently. The diagnostician provides structured reasoning but the team should validate against real user behavior.
• Factor ratings depend on evidence quality. Quantitative data (analytics, funnels, A/B tests) produces more reliable ratings than qualitative descriptions alone. The diagnostician classifies evidence quality and flags low-evidence assessments.
• Motivation is the hardest factor to assess remotely. Motivation involves internal psychological states that users may not articulate. Motivation assessments carry inherently higher uncertainty than ability or prompt assessments.
• Intervention effectiveness is context-dependent. All recommendations are directional, not guaranteed. Always validate through user testing or controlled experiments before committing to implementation.
• The Fogg model simplifies a complex reality. B=MAP reduces behavior to three factors for analytical tractability. Real behavior is influenced by additional factors (habit strength, environmental context, emotional state) not explicitly captured (Fogg, 2009).

---

Registration

This sub-skill follows a parent-routed registration model per H-26. Sub-skills are routed through the parent /user-experience orchestrator; independent registration would create duplicate triggers (AP-02).

Registration Point	Status	Detail
CLAUDE.md skill table	Registered via parent	/user-experience registered; sub-skills not independently listed
mandatory-skill-usage.md trigger map	Routed via parent	"behavior design" keyword routes to parent, which dispatches here
AGENTS.md agent registry	Registered	ux-behavior-diagnostician listed under User-Experience Skill Agents
Parent SKILL.md agent table	Registered	Listed in skills/user-experience/SKILL.md [Available Agents]

---

Deployment Status

Wave 4 Sub-Skill -- Phase 1 Complete, Phase 2 Pending.

This sub-skill follows a two-phase implementation sequence:

• Wave 4 Phase 1 (this deliverable): SKILL.md specification -- methodology, output format, routing integration, template stub, cross-framework integration, and quality gate criteria. This document is the Phase 1 artifact.
• Wave 4 Phase 2 (pending): Agent implementation -- skills/ux-behavior-design/agents/ux-behavior-diagnostician.md (agent definition with <input>, <capabilities>, <methodology>, <output> sections) and skills/ux-behavior-design/agents/ux-behavior-diagnostician.governance.yaml (governance metadata). Tracked under PROJ-022 EPIC-004.

---

Quick Reference

Common Workflows

Need	Command Example
Diagnose checkout abandonment	"Why aren't users completing the checkout process?"
Analyze onboarding drop-off	"Diagnose the behavioral bottleneck in our onboarding flow"
Understand CTA inaction	"Users see the upgrade button but don't click it -- what's blocking them?"
B=MAP factor analysis	"Analyze motivation, ability, and prompts for the signup action"
Post-launch behavior diagnosis	"Users are abandoning the form after step 3 -- diagnose why"
CRISIS behavioral diagnosis	"CRISIS: users are abandoning checkout -- urgent UX triage" (routes to full CRISIS sequence)

Agent Selection Hints

Keywords	Routes To
behavior design, B=MAP, Fogg model, behavior bottleneck, motivation analysis, ability analysis, prompt design, user inaction, action threshold, tiny habits	ux-behavior-diagnostician
heuristic, usability, Nielsen, severity, inspection, evaluation	/ux-heuristic-eval (not this sub-skill)
HEART, metrics, measurement, GSM, happiness, engagement	/ux-heart-metrics (not this sub-skill)
lean UX, hypothesis, assumption, experiment, build-measure-learn	/ux-lean-ux (not this sub-skill)
jobs to be done, JTBD, switch interview, user motivation	/ux-jtbd (not this sub-skill)
Kano, feature prioritization, must-be, attractive, performance	/ux-kano-model (not this sub-skill)

---

References

Source	Content	Path
Parent SKILL.md	Sub-skill scope, wave architecture, routing, MCP dependencies, synthesis protocol	skills/user-experience/SKILL.md
Agent definition	Agent frontmatter, identity, expertise, guardrails	skills/ux-behavior-design/agents/ux-behavior-diagnostician.md [PLANNED]
Agent governance	Tool tier, forbidden actions, output validation, constitutional compliance	skills/ux-behavior-design/agents/ux-behavior-diagnostician.governance.yaml [PLANNED]
UX routing rules	Lifecycle-stage routing, handoff data contracts, common intent resolution, CRISIS routing [PARTIAL: EPIC-001]	skills/user-experience/rules/ux-routing-rules.md
Synthesis validation	Confidence gate protocol, per-sub-skill confidence map, signal extraction criteria	skills/user-experience/rules/synthesis-validation.md [STUB: EPIC-001]
Wave progression	Wave 4 entry criteria, signoff requirements	skills/user-experience/rules/wave-progression.md
CI checks	P-003 enforcement, sub-skill validation gates	skills/user-experience/rules/ci-checks.md
B=MAP diagnosis template	Factor assessment tables, bottleneck algorithm trace, intervention format	skills/ux-behavior-design/templates/bmap-diagnosis-template.md
Skill standards	H-25/H-26 skill structure requirements	.context/rules/skill-standards.md
Agent development standards	H-34 dual-file architecture, tool tiers, handoff protocol	.context/rules/agent-development-standards.md
Quality enforcement	Quality gate thresholds, criticality levels, strategy catalog	.context/rules/quality-enforcement.md
Handoff schema	Canonical handoff schema v2	docs/schemas/handoff-v2.schema.json
Agent governance schema	Governance YAML validation schema	docs/schemas/agent-governance-v1.schema.json

Requirements Traceability

Source	Content	Path
PROJ-022 PLAN.md	Project plan: sub-skill scope, wave assignment, acceptance criteria, implementation phases	projects/PROJ-022-user-experience-skill/PLAN.md
EPIC-004 (Wave 4 Deployment)	Parent work item for Wave 4 sub-skill implementation including this sub-skill	PROJ-022 EPIC-004 in projects/PROJ-022-user-experience-skill/WORKTRACKER.md
ORCHESTRATION.yaml	Orchestration plan governing the build sequence for this sub-skill	projects/PROJ-022-user-experience-skill/orchestration/ux-skill-build-20260303-001/ORCHESTRATION.yaml

External References

Source	Citation
Fogg, B.J. (2009)	"A Behavior Model for Persuasive Design." Proceedings of the 4th International Conference on Persuasive Technology (Persuasive '09). Article No. 40. DOI: 10.1145/1541948.1541999. Original publication of B=MAP (then B=MAT). Defines motivator pairs, six simplicity factors, and three prompt types.
Fogg, B.J. (2020)	"Tiny Habits: The Small Changes That Change Everything." Houghton Mifflin Harcourt. Chapter 3: updated B=MAP with "Prompt" replacing "Trigger" and action line threshold model; Chapters 5-6: intervention difficulty gradient (Starter Step → Scaled Habit); Chapters 14-15: behavior statement format ("After I [anchor], I will [tiny behavior]") and scaling methodology.
Eyal, N. (2014)	"Hooked: How to Build Habit-Forming Products." Portfolio/Penguin. Complementary habit formation framework (Hook Model). Referenced for context, not directly applied.
Wendel, S. (2020)	"Designing for Behavior Change: Applying Psychology and Behavioral Economics." 2nd ed. O'Reilly. Chapters 5-7: practical design patterns for behavior change interventions targeting ability barriers (simplification strategies) and motivation barriers (framing, social proof, commitment devices). Referenced for intervention design patterns.
Fogg, B.J. (n.d.)	"Fogg Behavior Model" (living reference, actively maintained). https://behaviormodel.org/ (accessed 2026-03-04). Canonical online resource for the B=MAP model, updated factor definitions, and current practitioner guidance.

---

Sub-Skill Version: 1.5.0
Parent Skill: /user-experience v1.0.0
Constitutional Compliance: Jerry Constitution v1.0
Wave: 4 (Advanced Analytics)
SSOT: skills/user-experience/SKILL.md
Project: PROJ-022 User Experience Skill
Created: 2026-03-04